[Paper Reading] FaceTracer: A Search Engine for Large

Paper Site: https://www1.cs.columbia.edu/CAVE/publications/pdfs/Kumar_ECCV08.pdf

Problem Definition

Using a image search engine based entirely on face, search through a large scale database automatically labeled on the basis of several facial attributes with simple text queries.

Contribution and Discussion

  1. Created the first face search engine allowing users to search through large collections of images which have been autmatically labeled.

  2. Using a novel combination of Support Vector Machines and Adaboost shows srate-of-the-art classification results compared to previous works.

  3. Fully automatic, easy to scale and computes all labels off-line, leading to fast on-line search performance.

Methods

  1. Build a large database of over 3.1 million face images extracted from over 6.2 million images collected from the internet.
Overview of database creation
  1. Create a rich set of local feature options from which the classifier can automatically select the best ones: the region of the face to extract features from, the type of pixel data to use, the kind of normalization to apply to the data, and finally, the level of aggregation to use.
The face regions used for automatic feature selection
  1. Combine the strengths of these two methods by constructing a number of “local” SVMs and letting Adaboost create an optimal classifier using a linear combination of them. We create one SVM for each region, feature type, and SVM parameter combination, using the LibSVM library.
?著作權歸作者所有,轉載或內容合作請聯系作者
平臺聲明:文章內容(如有圖片或視頻亦包括在內)由作者上傳并發布,文章內容僅代表作者本人觀點,簡書系信息發布平臺,僅提供信息存儲服務。